374 DISCOVERY REPORTS 



on average to a similar fixed percentage of their original size. Presumably this is determined by the 

 initial size of the graafian follicle and the connective tissue framework, and it is conceivable that the 

 corpora of ovulation might regress to a different percentage of the initial size, owing to a possible 

 initial difference in the amount of connective tissue in the gland. 



We have material which enables us to check this hypothesis. It has been shown that the 'old' 

 corpora albicantia represent the final products of regression of corpora lutea both of ovulation and 

 pregnancy, and we have records of the size frequency of 2339 'old' corpora albicantia which enables 

 a very accurate frequency curve to be drawn. One assumption must be made — that corpora lutea 

 of ovulation and pregnancy are formed in the approximate ratio 2:1, but we can also test the hypo- 

 thesis using other ratios. There are good reasons, given later in this paper (see pp. 459-63), for 

 believing that this assumed ratio is in fact correct. This means that the mean diameter of all corpora 

 lutea will be 9-33 cm. and the upper regression line in Text-fig. 19 must be amended to the lower 

 for the fin whale (F x ) at least. 



In addition to the size frequency distribution of ' old ' corpora albicantia there are records of the 

 size frequency of 523 corpora lutea of pregnancy (Text-fig. 7). By using the graph (Text-fig. 19) these 

 size frequencies have been converted into the corresponding size frequencies for the resulting corpora 

 albicantia (Text-fig. 20, curve B). If both types of corpora regress to the same extent then the 

 frequency curve for all ' old ' corpora albicantia represents the sum of the frequency curves of ' old ' 

 corpora albicantia derived from corpora lutea both of ovulation and of pregnancy. 



For direct comparison these frequencies are converted into percentages. The intervals for the 

 calculated corpora albicantia of pregnancy are different from the intervals for total 'old' corpora 

 albicantia because the peak value of this curve is almost exactly 2-5 cm. If it were drawn to the same 

 intervals as the total 'old' corpora albicantia the result would be to convert a symmetrical curve to 

 a skewed curve, so the points are displaced by 0-25 cm. with reference to the total ' old ' corpora 

 albicantia. The true frequency curve for corpora lutea of pregnancy (Text-fig. 7) has a shoulder at a 

 size corresponding to the interval 2-0-2-5 cm. in corpora albicantia and this has been drawn in to make 

 the curve more symmetrical (Text-fig. 20, curve B). 



If the old corpora albicantia curve is the sum of fully regressed corpora lutea of ovulation and 

 pregnancy in the ratio 2:1, it may be considered to be composed of three units, and the corpora 

 albicantia of ovulation and pregnancy as two and one units respectively. In Text-fig. 20 the corpora 

 albicantia of pregnancy (B) total 100% and total 'old' corpora albicantia (A) are converted to equal 

 300%. If our assumption is correct a symmetrical curve totalling 200% (C) should be obtained by 

 subtracting the corpora albicantia of pregnancy frequencies (B) from the total ' old ' corpora albicantia 

 frequencies (A). This curve (C) should have a mean, mode and size range approximating to these 

 values for the sample of corpora lutea of ovulation sizes converted to corpora albicantia of ovulation 

 sizes. In Text-fig. 20 the size range, mean value and two standard errors (D) for the calculated 

 corpora albicantia of ovulation are shown for comparison because the sample is too small to give a 

 smooth curve. In fact, curve C is symmetrical, and its mean and mode approximate very closely to 

 the mean corpus luteum of ovulation diameter converted to corpus albicans of ovulation size. 



The hypothesis can be tested with other ratios of corpora lutea of ovulation and pregnancy, but 

 none gives such a good fit. For instance, on the assumption that the numbers of corpora lutea of 

 ovulation and corpora lutea of pregnancy are equal, curve A is drawn so as to total 200 %, and curve B 

 as 100%, so that curve C should equal 100%. Curve C is again symmetrical with apeak at i-8 cm., 

 but the size range is much less than expected (from D) and part of curve B falls outside curve A, 

 which is improbable. 



It is, therefore, probable that the ' old ' corpora albicantia represent fully regressed corpora lutea of 



